“ Lets clear up one thing - OCR (Optical Character Recognition) software is but one means to create a text layer. Additional software is used to embed that text layer into file types that accept a predefined standard for text-layers such as PDFs and DjVus. Expensive OCR programs have numerous predefined routines based on tedious trial and error over time in addition to the ability to customize paramaters for near perfect results. The freeware-ish stuff we are exposed to online is severly less than that. ”

“ We all here work for a better world. The ascii is one part, the primary portion, and the code yet another part of a process. In short, we who do these combinations of things make a better work when combining these processes for everyone, blind and deaf and not blind nor deaf. Code makes the book look pretty—but not for the blind it doesn't. ”

“ Well it was BeesWax who actually modified the Pagelist's page numbering/designations after I had pointed out that "proper" scan page verification "should" always demand that this be done prior to any actual proofreading. Without it being in place, we frequently find ourselves in situations where duplicated/omitted/corrupted individual pages are not discovered until well into the proofreading regime. This means that dozens, if not hundreds, of newly created pages have to be moved in one manner or another to properly align with a newly corrected or fully substituted DjVu file. ”

“ Hi George, I was trying to do some editing, updating executive orders by POTUS using your POTUS-eo template, TextQuality template and Font-size template but got a lot of errors. Can you possibly visit my sandbox and check what's in there, if that's even possible? I'm new to wiki editing and trying to get the hang of everything. The info I have in the sandbox for the executive order page creation is valid, and I wanted to follow all the great contributions you have already made to the topic, but the formatting doesn't cooperate. ”

“ The problem is the nearly same when forcing all-to-the-left only a bit worse since side-notes and page links back to the Page: namespace can, and frequently do, overlap each other thanks to the Dynamic layout settings in .js & .css. What you're trying to accomplish, I think, is keeping the pagenum links all the way to the left up against the navigation, toolbox, etc. sidebar like you'd normally see in Layout 1, then have have all the side-notes fall in between those links and the end of the left margin of the text-container box under Layout 2. ”

“ The most annoying and "least solved" by the further development of the above to its theoretical end concerns images. Images do not shrink and grow in relation not only to the page size (or view screen) in use, but to any of the other content that the image may share a page with. ”

“ IMO, we're not "stuck"; we accept realities without measuring the consequences because it "works" for most of the regular contributors most of the time and in most instances for the majority of works hosted to date; case closed. Look... using straight HTML for tables allows the pagenum class to render absolute across page breaks in the mainspace because the span has somewhere to "land" properly (i.e. between a and once transcluded) . The same principle (s) for pagenum within a table across a page-break applies to a sidenote as well ”

“ I suppose it could be done in the same general manner as the page numbers, by using Javascript, but I don't have the energy or time any more to fight that kind of change through the system. An alternative is to put the sidenotes in the other column of the table, but that breaks the desirable orthogonality between table layout and sidenote handling (they should not affect each other) , and it makes maintenance harder if the sidenotes are "flipped". I understand all this. Layout 1 will never provide any margin for sidenotes because the 3 div wrappers have no styling defined in the Common.js. ”

“ The mainspace status bar is missing the "white" color normally know as the indicator a Page: has been allocated in the pagelist tag but has not being created yet. As it is now, the status bar would show all green for the six pages that exist and the vistor would have no clue 4 pages are missing (your lopsided proportioning) . That's an important fault (imo) . It shouldn't be rigged. I should be able to tell if only 10% of the book is transcribed (a common situation) . ”

“ A typesetter may also make mistakes for a scanned .PDF or DJVU book book whereas you may have submitted the work properly. Letters, as you probably know, were pressed down many times to where an h can look like a b or a c can look like an o. Typesetters do not always throw a type back into the melting pot and make a new letter so as not to make mistakes. I worked, as did another around here, as a type setter for a summer so I know that process from old books. Sure, you can read an imperfect .PDF file but the blind cannot. ”

“ The problem is that several template variations (sidenotes outside or all left-to-right, etc.) strip the span class for sidenotes in the mainspace (making dynamic control over the span class completely moot) or strips the wrapping span tag altogether leaving no place for the sidenote to "land" in transclusion (other than before or after the entire table; not the table cell) . I remain opposed to deletion unless the God damn implementation of the God damn selectable first-click-in of a God damn dynamic layout for a God damn mainspace work is finally made permanent. ”

“ Inline wikilinking within article content is always the norm; not the exception. -- George Orwell III (talk) 21:39, 22 June 2012 (UTC) I do understand the viewpoint (only links to works of the author) and will abide by it in the future; but last week I simply took a different point of view: in the spirit of a wiki I wanted to link what I saw as a "unique pointer" (in contrast to the "Hood" in you example, which would be senseless as a link and not have a page here) . ”

“ You won't find much info on this because DjVu is not a popular file format being used by the virtual masses out there - PDFs are, and that whole other ball of wax - so nobody has made a (free) way to "do" what we're missing (yet; somebody needs to develop the rest is all) . Also understand that the above complexities can be achieved with PDF files but that currently involves lots of licensed software ($$$) so we don't have-it/develop-it/use-it in the Wiki-world as other entities might. ”

“ Clarify: look for a text-layer to pop into the edit window when you go to create a page for the first time under Edit mode (not via the OCR button) .When text appears in the edit box automatically, it is the embedded text-layer dumped from the source file. When you hit the OCR button, a program tries to detect the text from the thumbnail of a scanned page and dumps the results into the edit box. ”

“ As a starting point, things seem generally to center fine in Mobile view for centering of text (can you see this?) . Thumbnail images, however (see same link) , no matter how you format their alignment, seem to automatically center no matter what. The only thing that doesn't seem to center in Mobile view are regular images (i.e., not thumbs) with the centering 'command' within the transclusion formatting. Right alignment (within image transclusion brackets) seems to work, however. And centering images works if you "force"-center the image with the centering formatting outside the brackets. ”

“ If say a "double-entry" exists, I've seen this throw off the layer to thumbnail marriages before. You'd then need to use DjVused.exe to properly extract, repair and merge the outline in the DjVu file. I could not find your command. What I did is looking into: DjView->View->Information but could not find any double or missing entry in the sequence. If you extracted the text-layer to a .txt file using DjVused.exe, run a 'find' against it using page 0 0 (thats the word page, space, zero, space, zero, space) . ”

“ There is no "neat" way to convert the old template to account for these new factors either & if there is a better way to construct their URL - the folks at FDsys haven't shared it with me yet (if ever) .
What I did manage to do, as you've most likely already figured out, is auto create an FDsys URL that opens not-so-optimal PDF version of the page (s) wanted in the section entered. This URL only required a rather minor template mod where the E, H, S, D section designator is split off and is it's own independent parameter from the page number input. ”

“ On the one hand Layout 1 does not impose a fixed width nor absolute:relative positioning but you can't "see" the side-notes anyway in anything other than some floating box (i.e. no margins) . This is not the case under Dynamic Layouts 2 and 3 only by chance because the idea is to force to the right where a template can "overflow" into the screen-space regardless of the widths and margins imposed by Dynamic Layouts (its whitespace to begin with in other words) . ”

“ This Java issue requires a command line parameter to increase the memory stack but I haven't figured out which java .exe is called by pdfsam and how to add the parameter. The one thing I haven't figured out is the software you use to remove watermarks, and whether this is done manually page by page, or there is an automatic way to remove all watermarks. If done manually, page by page, then I know the true profitability of the undertaking. Currently, there are three uploads on IA waiting for processing, but I don't expect them until after Labour Day. ”

“ All of that mobility is wasted in en.WS because the PR extension only deals with 'bundled' DjVu files - files where there is no central 'index' DjVu file holding the page progression and other such data. This info is spread across all the individual DjVu files that make up a 'bundled' DjVu file - making it rather hard to manipulate the addition, deletion and replacement of one or more individual files.
Second thing to keep in mind is not much can be achieved on the fly or on-line. ”

“ The above scroll-box is the "view-source" -like resulting output of executing one of the other files that comes with the DjVuLibre software package (DjVutoXML.exe) against any given .DjVu source file. It is meant to illustrate the difference in complexity compared to the output of DjVuDump.exe (or DjVuTxt.exe? I forget) which are also DjVuLibre packaged software pieces that are currently installed on our servers and used to "dump" the embedded text-layer to the Page: namespace when you go to create a page for the first time. ”

“ If some proofreading of an Index has already been done, would such text layer recovery affect the already-proofread pages? (My line of questioning is starting to be based on assumptions, so I'll stop here) Londonjackbooks (talk) 21:55, 30 November 2012 (UTC) Not at all. Since developers early on opted for a plain old read (the source-file text-layer) and dump (it as plain text to the Page: namespace) method, nothing you add, edit or save text-wise will ever change the text-layer in the source file. ”

“ This is usually not display affecting for most of us and OK to do more often than not because we, in general, are addicted to repetative template application for the slightest of changes of the most common of attributes given to the absolute basic formatting teg elements. So we wind up wraping content in container upon container upon container until things display well enough for our satisfaction instead of letting the standards do the lifting they were crafted to do by default on their own. ”

“ Sounds similar to the 'Xerox of a Xerox of a Xerox of a Xerox of type written page of text paradox' no? Downloaded the volume and stored it in the program directory. Using djvm.exe in command line mode, deleted the two damaged pages containing the fingers of the IA volunteers, and then verified the results with djvu in GUI mode (very fast) . Merged the two new (consecutive) pages using djvm.exe in command line mode, into one file and verified the results with djvu in GUI mode before proceeding to the next step. ”

“ We'll be hacking, designing, teaching, and socialising, primarily talking about ResourceLoader and Gadgets (extending functionality with JavaScript) , the switch to Lua for templates, Wikidata, and Wikimedia Labs.
We want to bring 100-150 people together, including lots of people who have not attended such events before. User scripts, gadgets, API use, Toolserver, Wikimedia Labs, mobile, structured data, templates -- if you are into any of these things, we want you to come! ”

“ Technology is always "upgrading" and I like to upgrade things I like. If I want, I can add hyperlinks to the .djvu book now being worked on but I doubt (at the moment) I will even bother with that. Those scientists are amazing people though. My original intent was to create an illustrated version of the book that has been on WS for the past 3 years. ”

“ The remaining .js files you mentioned should be copied not imported to each sub-domain. The further development of said files then becomes exponentially more exposed to possible imptovements because more local folks are able to access and change the file.Right now, the few who have bothered to change/improve things are given access privelages instead to "market" there own ideas across one or more wikis they never normally else would contribute to material works wise. I'm not some meanie who wants to block useful folks from making improvements beneficial to many or all domains ”